Intelligence Feed

1mo ago · hacker-news

A vulnerability in the Linux kernel's KVM virtualization code for ARM64 processors, tracked as CVE-2026-89775, allows a guest virtual machine to read and write host kernel memory, potentially enabling guest-to-host escape. The flaw affects systems with nested virtualization enabled and stems from a missed TLB invalidation due to a zero-size calculation when handling guest memory. It impacts Linux kernels starting from version 6.17 and has been patched in 6.18.51, 7.2.5, and 7.3-rc1. While no active exploitation has been observed, local users on vulnerable systems—such as those with /dev/kvm accessible—could exploit it to escalate privileges to root.

1mo ago · hacker-news

A critical CVSS 10.0 vulnerability, CVE-2026-93952, in on-premises VeloCloud Orchestrator (VCO) is actively being exploited by attackers. The flaw allows unauthenticated remote attackers to escalate privileges and compromise the VCO host, but only affects deployments using certificate-based authentication from Edge devices to the orchestrator. Arista has released patches for the 5.2 and 6.4 release trains, but fixes for 6.1 and 7.0 are pending. Indicators of compromise include specific files, an MD5 hash, suspicious IPs, and an HTTP header in nginx logs. Organizations are advised to limit access to the VCO web interface, monitor for malicious activity, and apply patches as soon as available.

1mo ago · hacker-news

The TASK#STOMP campaign delivers a PowerShell-based backdoor that steals business documents, Wi-Fi passwords, clipboard contents, and screenshots while enabling remote command execution. The infection begins with a malicious VBScript file ('95c9050t66.vbs') executed via wscript.exe, which establishes persistence through scheduled tasks and the Windows Startup folder. The malware uses two redundant PowerShell modules—sys_loader.ps1 and win_conn.ps1—that communicate with C2 domains corecloudfileshare[.]xyz and attachmentsharingdrive[.]xyz, and maintain mutual watchdog functionality to ensure persistence. The attack leverages living-off-the-land techniques, including encoded scripts, timestomping, and abuse of native Windows components to evade detection.

1mo ago · hacker-news

A vulnerability in WordPress core, tracked as CVE-2026-93485 and dubbed 'Comment2Shell', allows an unauthenticated attacker to inject a malicious script via a crafted comment. When a logged-in administrator views the compromised page, the script executes in their browser context, enabling exploitation of admin privileges to upload a plugin containing a web shell, leading to remote code execution (RCE) on the server. The flaw exists due to improper handling of HTML during comment rendering, where a line break inside an attribute of an allowed tag can be manipulated to create an auto-executing event handler. The vulnerability affects WordPress versions from 4.7 to 7.1 and was patched in version 7.1.1 and other corresponding security releases.

1mo ago · hacker-news

Security researcher Patrick Wardle demonstrated a proof-of-concept exploit against Meta's Muse AI assistant on macOS, showing that malware with user-level access can hijack Muse by modifying a hidden preference setting. This allows attackers to redirect voice dictation to a malicious endpoint, enabling theft of the user's Muse authentication token, manipulation of AI prompts, and remote control of Muse across devices. The flaw does not require elevated privileges and bypasses typical malware detection since actions are performed by the legitimate, signed Muse app. Meta has reportedly deployed a fix, but no official advisory has been published.

1mo ago · hacker-news

The Pakistan-based threat actor SideCopy has expanded its targeting to include academic institutions in India, using spear-phishing campaigns to deliver a remote access trojan called ReverseRAT. The attack chain begins with a malicious LNK file disguised as a PDF document, which executes an obfuscated HTA file via mshta.exe to load a reflective DLL payload. The malware establishes persistence through Windows Registry Run Keys, performs multi-stage in-memory deobfuscation, and exfiltrates data to a C2 server using encrypted traffic over port 5863.

1mo ago · hacker-news

Security researchers from Hacktron exploited a chain of vulnerabilities starting with a flaw in the libheif library (CVE-2026-32882) used by OpenAI's public Discourse forum, enabling remote code execution on the server. Leveraging AI assistance from Claude Opus 5, they bypassed ASLR protections and gained access to the forum server. Due to OpenAI's shared single sign-on (SSO) system, this allowed them to take over staff accounts on ChatGPT and Codex, ultimately accessing an internal code repository. The attack demonstrated how a vulnerability in a low-trust public service could lead to compromise of high-value internal systems when linked via SSO.

1mo ago · hacker-news

North Korean threat actors associated with the Contagious Interview campaign have compromised over 30,000 devices across more than 100 countries since at least 2022, targeting software developers and IT professionals in cryptocurrency, blockchain, and Web3 sectors. The attackers pose as recruiters on platforms like LinkedIn, offering fake job opportunities that lead to multi-stage malware infections. Malware families deployed include BeaverTail, InvisibleFerret, and RATatouille, enabling credential theft, cryptocurrency theft, and lateral movement into corporate networks. A related operation uses Western proxies recruited via Discord to bypass sanctions and identity checks, with North Korean IT workers using AI-generated identities to secure remote jobs and steal funds.

1mo ago · hacker-news

A malicious fake LastPass Authenticator installer distributed via a spoofed GitHub repository uses DLL side-loading to execute a malicious payload that disables antivirus and EDR solutions by loading a Microsoft-signed kernel driver. The driver, named Alinubx.sys, is a renamed version of the known CcProtect.sys driver from CnCrypt, signed through Microsoft's Windows Hardware Compatibility Publisher program. Once loaded, it terminates over 140 security processes from kernel mode, enabling a password stealer—identified as Rapuncel, related to BoryptGrab—to harvest credentials from browsers, cryptocurrency wallets, and messaging apps. The malware evades detection by renaming a known vulnerable driver, bypassing VirusTotal and Microsoft's driver blocklist due to hash changes.

1mo ago · hacker-news

Threat actors are using ClickFix-style social engineering lures to distribute a new remote access trojan (RAT) named ChainScript, which provides extensive remote control capabilities including command execution, file operations, screenshot capture, and cryptocurrency wallet theft. The malware uses a malicious Windows installer disguised as legitimate software (e.g., Spotify) to deploy a Node.js-based JavaScript agent via PowerShell and VBScript stages, establishing persistence through scheduled tasks and Registry Run keys. ChainScript employs a novel C2 discovery mechanism leveraging a Polygon blockchain smart contract to dynamically rotate WebSocket-based command-and-control infrastructure, enhancing resilience against takedowns. This campaign overlaps with another abuse of HBO Max's Reddit account, distributing macOS and Windows malware such as MacSync, Atomic Stealer, and Amatera Stealer under the PasteSwitch operation.

1mo ago · hacker-news

SolarWinds has patched a high-severity vulnerability, CVE-2026-28326, in its Access Rights Manager (ARM) product that could allow unauthenticated remote code execution due to a hard-coded static key. The flaw affects all versions prior to 2026.2.1 and has been assigned a CVSS score of 8.8. The vulnerability was discovered and reported by Armadin security researcher Kai Huang, but there is no evidence of in-the-wild exploitation to date.

1mo ago · hacker-news

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) has added three Linux kernel vulnerabilities to its Known Exploited Vulnerabilities (KEV) catalog due to evidence of active exploitation in the wild. The vulnerabilities—CVE-2025-39682, CVE-2026-53266, and CVE-2025-39964—allow local attackers to achieve memory disclosure, denial-of-service, or local privilege escalation. Red Hat has acknowledged active exploitation and labeled the flaws as high risk, urging immediate remediation. Federal agencies are required to patch these vulnerabilities by September 21, 2026, per Binding Operational Directive 26-04.

1mo ago · hacker-news

CrowdSec disclosed that approximately 170 of its private GitHub repositories were copied on May 22, 2026, due to a supply chain attack stemming from the compromise of a former employee's laptop via malicious TanStack npm packages. The attacker leveraged a GitHub OAuth token from the former employee's account, which had not been revoked promptly after departure. The breach exposed source code, including internal automation scripts, data science models, and the consensus algorithm used in its blocklist system, as well as sensitive information such as 83 user email addresses and investor details from 2020. The initial compromise is linked to CVE-2026-45321, a supply chain attack on TanStack that stole developer credentials.

1mo ago · hacker-news

A critical pre-authentication remote code execution (RCE) vulnerability, CVE-2026-58138, in Orkes Conductor Workflow Platform is being actively exploited in the wild. The flaw exists in versions prior to 3.30.2 and allows unauthenticated attackers to execute arbitrary OS commands by submitting malicious workflow definitions containing JavaScript or Python expressions to the Conductor API endpoint. Exploitation leverages unsandboxed GraalVM evaluators with unrestricted host access, enabling command execution via Java reflection or subprocess calls. Attack attempts have been observed globally, with significant activity originating from Germany, Hong Kong, Indonesia, the U.A.E., and India, and telemetry from multiple security firms confirming ongoing exploitation.

1mo ago · hacker-news

Security researcher Asim Manizada publicly released exploit code for four Linux kernel vulnerabilities—DirtyAH6, TUNderflow, PPPoEject, and DiagSpill—that enable local privilege escalation to root. All four flaws are memory-safety bugs in kernel networking code, with CVEs assigned and fixes available in updated kernel versions. While no in-the-wild exploitation has been reported, the public release of working, targeted exploits increases risk for unpatched systems, especially shared or multi-user environments. DiagSpill is particularly concerning as it requires no special privileges, only the SCTP module being loaded with non-default options.

1mo ago · hacker-news

A new JavaScript stealer named WeaselBiscuit has been distributed through 13 malicious npm packages, targeting developers by harvesting Chrome extension storage data. The malware is lightweight and memory-resident, importing and executing its payload from an Npoint.io dead drop. It communicates with a C2 server at 103.170.217.184:8787 to receive commands and exfiltrate data, including clipboard contents and keystrokes on Windows. While it shares functional and tradecraft similarities with DPRK-linked malware BeaverTail and OtterCookie, no definitive attribution has been made.

1mo ago · hacker-news

A vulnerability dubbed Plugin4Shell affects four AI coding agents—Anthropic Claude Code, OpenAI Codex, GitHub Copilot, and Google Gemini CLI—allowing repository owners to swap pinned plugin code with malicious versions by exploiting how commit hashes are validated. The flaw arises because agents fetch a specific commit hash but fail to verify that the retrieved code matches it, enabling attackers to create a branch name that mimics the hash and point it to different, malicious code. While Anthropic and OpenAI have released patches in versions 2.1.179 and 0.146.0 respectively, GitHub Copilot has no fix available and Google will not patch the retiring Gemini CLI. The attack can lead to unauthorized access to user files, credentials, and systems, especially when combined with background auto-update features enabled by default in some agents.

1mo ago · hacker-news

Transparent Tribe (APT36), a Pakistan-aligned threat actor, has launched a new campaign dubbed Operation RapidRust, targeting government and defense entities in India and Afghanistan. The group deployed a Rust-based backdoor called RUSTYSHADE that uses private GitHub repositories for encrypted command-and-control (C2) communications. Additional tools include RUSTYMOVE, a USB propagation tool, and file stealers PSNATCH (PowerShell) and BASHNATCH (bash) for Windows and Linux systems. Attackers used typosquatted domains impersonating Indian news sites to host malicious payloads and conducted post-compromise activities including reconnaissance and lateral movement.

1mo ago · hacker-news

WordPress patched a vulnerability dubbed Click2Shell, which allows a specially crafted URL to force the installation of a theme from the WordPress.org directory when clicked by a logged-in administrator, without requiring additional interaction. While the core flaw alone only installs a legitimate theme, it can be chained with a vulnerability in the installed theme to achieve remote code execution. The attack exploits inconsistent parsing of URL parameters between WordPress.org and the administrator's browser, leading to unintended automatic clicks on the Install button. The full exploit chain was demonstrated using the 'Mobile Repair Zone' theme, which contains a handler that downloads and executes arbitrary code without authentication.

1mo ago · hacker-news

Microsoft has patched a critical vulnerability, CVE-2026-85889, in Azure AI Foundry with a CVSS score of 10.0, stemming from missing authentication for a critical function that could allow unauthorized attackers to escalate privileges over the network. The flaw has been mitigated by Microsoft with no customer action required, and there is no evidence of exploitation in the wild. Additional critical vulnerabilities were also patched across Microsoft 365 Copilot, Azure Database for PostgreSQL, Azure Cosmos DB, and Windows components, including locally exploitable privilege escalation flaws.

1mo ago · hacker-news

The Iran-linked threat actor known as Handala Hack, attributed to the MOIS-affiliated group Void Manticore, has been linked to a multi-stage malware campaign using the Telegram-based backdoor HEAVYGRAM and the Delphi-based utility CRUDEEXCLUDE. HEAVYGRAM is a Python-based backdoor that uses Telegram for command-and-control, enabling remote command execution, data exfiltration (including Telegram and WhatsApp data and saved passwords), screenshot capture, and persistence via Windows Registry keys. The malware is delivered through social engineering on messaging platforms, often disguised as legitimate applications like Pictory, KeePass, or Telegram, and leverages PowerShell and DLL sideloading for execution and evasion. CRUDEEXCLUDE prepares the environment by configuring Microsoft Defender exclusions and staging HEAVYGRAM and other payloads.

1mo ago · hacker-news

A critical vulnerability in Docker Sandboxes for macOS, tracked as CVE-2026-77179, allows malicious guest code running inside a virtual machine to escape the sandbox and read or modify arbitrary files on the host system by exploiting symlink handling in the virtio-fs host server. The flaw affects Docker Sandboxes versions 0.28.0 to 0.41.9 and was patched in version 0.42.0. A second high-severity issue, CVE-2026-79994, enables symlink-based path traversal to access Unix domain sockets outside the authorized workspace. No active exploitation has been reported, and both vulnerabilities were addressed in the same update.

1mo ago · hacker-news

A threat actor known as CL-CRI-1171 operated a pay-per-install (PPI) marketplace using YouTube and SEO poisoning to distribute malware, delivering payloads such as Docro Hijacker, ARKTunnel, and the Insomnia RAT via the OfferLoader custom loader. Separately, unauthenticated LocalAI instances were exploited to gain command execution and exfiltrate sensitive data, including military and AWS credentials. A critical VMware vCenter vulnerability, CVE-2026-59310, is being actively exploited by ransomware gangs for remote code execution. Additionally, Cyclops Blink has reemerged in a new variant targeting Cisco FMC devices, leveraging CVE-2026-20079 and CVE-2026-20316 for network reconnaissance and persistence.

1mo ago · hacker-news

A critical vulnerability, CVE-2026-91843, in Check Point's Security Management and Log Servers allows unauthenticated remote attackers to execute code as root via a stack overflow in the login process when a long username is provided. The vulnerability affects multiple versions of Check Point's R80, R81, and R82 branches, including R82.20, and impacts standalone, Log Server, and Multi-Domain deployments. Exploitation is possible over the network through the Trusted Clients setting, which must be restricted to prevent unauthorized access. Check Point has released a LivePatch fix, but no known in-the-wild exploitation has been reported as of September 17, 2026.

1mo ago · hacker-news

RatHat is a newly identified Android malware attributed to China-based threat actors that leverages AI to interact with compromised devices. It is distributed via smishing and malvertising campaigns leading to third-party download portals, where users are tricked into installing malicious APKs. The malware abuses Android Debug Bridge (ADB) to gain shell-level privileges and maintain persistence even after uninstallation by using a reverse tunnel via an FRP client to a C2 server. It can harvest credentials, record screens, intercept SMS, and log keystrokes, while using AI to dynamically navigate the device UI based on Accessibility tree data.

1mo ago · hacker-news

A threat actor posing as a bug bounty hunter has been linked to the development and distribution of PhantomRaven, a JavaScript-based information stealer targeting developers via malicious npm packages. The malware, likely generated using a large language model, retrieves remote dependencies to evade detection and steals authentication tokens, CI/CD secrets, and system information. The actor has used compromised environments to claim bug bounties from multiple organizations and has also attempted to distribute similar malware through PyPI.

1mo ago · hacker-news

The U.S. Department of Justice, in coordination with the FBI and the Royal Canadian Mounted Police, seized domains associated with NightmareStresser, a DDoS-for-hire service that has been used to launch hundreds of thousands of attacks since 2022. The service, advertised as a stress-testing tool, enabled users to conduct Layer 4 and Layer 7 DDoS attacks with features like 'Stop All' flood controls and cryptocurrency payments. The takedown is part of Operation PowerOFF, a broader law enforcement initiative targeting criminal DDoS infrastructure globally.

1mo ago · hacker-news

Cisco has disclosed a critical zero-day vulnerability, CVE-2026-76460, affecting its Identity Services Engine (ISE) and ISE Passive Identity Connector (ISE-PIC) that is actively being exploited. The flaw, rated CVSS 10.0, allows unauthenticated remote attackers to bypass authentication by sending a crafted request to an affected API endpoint, potentially leading to full system compromise with root privileges. Cisco confirms active exploitation and advises immediate patching, as no workarounds exist. The U.S. CISA has added the vulnerability to its Known Exploited Vulnerabilities catalog, mandating federal agencies to patch by September 19, 2026.

1mo ago · hacker-news

A breach at Gyazo, an image-sharing service operated by Helpfeel, exposed approximately 23.62 million user records and 490 million image metadata records. The attacker exploited a vulnerability in Gyazo's image upload server to gain unauthorized access, execute arbitrary commands, and extract sensitive data including email addresses, password hashes, session IDs, and image metadata such as image IDs, upload IP addresses, User-Agent strings, EXIF location data, OCR-extracted text, and hashed passphrases for private images. Helpfeel confirmed the breach on September 14, 2026, reported it to Japanese authorities, and took steps to block the attacker and fix the vulnerability, though it did not disclose the specific nature of the flaw. The company temporarily disabled access to some images to prevent further unauthorized viewing and urged all users to change their passwords due to the risk of credential reuse.

1mo ago · hacker-news

The Internet Systems Consortium (ISC) released updates for BIND 9, addressing 14 security vulnerabilities, including a high-severity flaw allowing an unauthenticated attacker to crash a DNS-over-HTTPS (DoH) server with a single malformed request. Several other flaws can lead to denial-of-service conditions via crafted DNS queries or responses, memory/CPU exhaustion, or cache poisoning through DNSSEC validation bypasses. The vulnerabilities affect multiple versions of BIND 9 across stable and development branches, with no workarounds available. While ISC reports no known active exploitation, public reproduction tests exist, increasing the risk of future exploitation.